Friday , September 25 2020
Home / ARTIFICIAL INTELLIGENCE / “The term AI is so sexy that it makes calculations look like intelligence”

“The term AI is so sexy that it makes calculations look like intelligence”

Tribune. He is a prominent figure in artificial intelligence (AI), Yann Le Cun, who points out that the prowess of AI demonstrates human intellectual limits much more than the intelligence of its achievements. We are bad calculators, we cannot brew large amounts of information, our memory is not reliable.

From this point of view, it is less surprising that AI boosted by gigahertz and terabytes crush human experts in exercises that we recently considered as intellectual heights. AI pioneer John McCarthy (1927-2011) would be annoyed – “As soon as it works, we stop calling it AI” (“As soon as it works, no one calls it AI any more”, according to the words reported by the polytechnician Bertrand Meyer on his blog. Does this mean that, for sixty years, there has been a misdeal on intelligence? What AI is chasing after a chimera? It would seem good …

So should we renounce the very designation of the discipline? I have no illusions: the term AI is so sexy, how can I give it up? I believe that it will continue for a long time to cause confusion in the public, the media and the industry, causing calculations to be taken for intelligence, algorithms for terminators.

Article reserved for our subscribers Read also “Work will not go away because of artificial intelligence”

The limits of AI show, at least, that his achievements are by no means artificial creatures. They do not understand what they are doing or who they are dealing with. They are not aware of the problems they deal with, the importance that humans attach to them, nor the benefits they hope for. They know nothing about the dreams of creatures of the flesh, of emotions, of filiations, of ambitions, of desires and of fears. How could algorithms grasp our problems, respond to the contradictory expectations that we find it difficult to formulate when we don't hide them under the veneer of good-thinking? How can we weigh the advantages and disadvantages, “discuss” them with us in order to reach compromises that we will undoubtedly challenge? Vast challenge for a machine…

Animal spirits

It may be useful to start by descending a few steps from the pedestal on which we climbed. We only perceive a tiny part of the events, because they are invisible, too fast or too slow, or too numerous. For those whom we manage to grasp, we have great difficulty in establishing what relates to cause or effect, lost by the incessant recombination of causality into cascades of infinitely varied effects. Our mind detects only the patterns that our limited senses can register, it detects coincidences, but on condition that they hit our senses intensely.

Leave a Reply

Your email address will not be published. Required fields are marked *